基于IMM滤波器的纯方位机动目标跟踪

IMM Filter with Application to Bearings-only Passive Maneuvering Target Tracking

针对无源纯方位跟踪中目标机动的问题,提出了一种基于交互式多模型的目标跟踪算法。该算法用伪量测变换估计器(PLE)将纯方位跟踪中非线性观测模型线性化,避免了计算雅克比行列式。机动目标跟踪中通过实时调整模型匹配概率,提高了滤波器对状态变化的跟踪能力。同时该算法实时修正观测噪声协方差,消除目标远离基阵时观测噪声对目标定位的影响。最后通过与MGEKF进行比较,Monte Carlo仿真结果验证了该算法的优越性。

A new IMM filter is presented for the problem of bearings-only passive maneuvering target tracking. Before the IMM filter, a pseudo-linear estimation （PLE） is used to restructure the nonlinear measurement equation, it has a brief form and little computation. The algorithm has strong robustness against model mismatching by updating the mode probability on-line, and it can avoid big error caused by searching inaccurate modified function and detecting maneuvering in Modified Gain EKP （MGEKF） algorithm. By adjusting the measurement covariance on-line, new IMM filter can eliminate the effect of measurement noise to target location when target far from sensor. At last, Monte Carlo simulation results show that this algorithm is better than MGEKF.